Performance Comparison of Random Forest and XGBoost for Diabetes Prediction

Abstract

Machine learning is widely used in modern healthcare systems to support disease prediction and diagnosis. Diabetes is one of the most common chronic diseases, and early prediction is important to reduce long-term risks. This paper compares two popular machine learning models, Random Forest and XGBoost, for diabetes prediction. A publicly available diabetes dataset is used, and the models are evaluated using accuracy, precision, recall, and F1-score. Experimental results show that XGBoost performs slightly better in accuracy and precision, while Random Forest performs competitively with simpler tuning and faster execution. The study concludes that both models are effective, but XGBoost provides better overall performance for healthcare prediction.